[ 
https://issues.apache.org/jira/browse/SOLR-7836?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14698671#comment-14698671
 ] 

Yonik Seeley commented on SOLR-7836:
------------------------------------

OK, I let the current unmodified trunk run overnight.
I got 14 fails, 4 of which were deadlocks.

Here's the only changes I made to the test:
{code}
Index: solr/core/src/test/org/apache/solr/search/TestReloadDeadlock.java
===================================================================
--- solr/core/src/test/org/apache/solr/search/TestReloadDeadlock.java   
(revision 1695727)
+++ solr/core/src/test/org/apache/solr/search/TestReloadDeadlock.java   
(working copy)
@@ -74,7 +74,7 @@
     int nWriteThreads = 5 + random().nextInt(10);
 
     // query variables
-    final AtomicLong reloads = new AtomicLong(50);  // number of reloads. 
Increase this number to force failure.
+    final AtomicLong reloads = new AtomicLong(200);  // number of reloads. 
Increase this number to force failure.
 
     ifVerbose("commitPercent", commitPercent, "deleteByQueryPercent", 
deleteByQueryPercent
         , "ndocs", ndocs, "nWriteThreads", nWriteThreads, "reloads", reloads);
@@ -177,7 +177,7 @@
 
     // The reload operation really doesn't need to happen from multiple 
threads, we just want it firing pretty often.
     while (reloads.get() > 0) {
-      Thread.sleep(10 + random().nextInt(250));
+      Thread.sleep(1 + random().nextInt(50));
       reloads.decrementAndGet();
       h.getCoreContainer().reload("collection1");
     }
{code}

> Possible deadlock when closing refcounted index writers.
> --------------------------------------------------------
>
>                 Key: SOLR-7836
>                 URL: https://issues.apache.org/jira/browse/SOLR-7836
>             Project: Solr
>          Issue Type: Bug
>            Reporter: Erick Erickson
>            Assignee: Erick Erickson
>             Fix For: Trunk, 5.4
>
>         Attachments: SOLR-7836-synch.patch, SOLR-7836.patch, SOLR-7836.patch, 
> SOLR-7836.patch, deadlock_3.res.zip, deadlock_5_pass_iw.res.zip, deadlock_test
>
>
> Preliminary patch for what looks like a possible race condition between 
> writerFree and pauseWriter in DefaultSorlCoreState.
> Looking for comments and/or why I'm completely missing the boat.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to